Supplementary materials for: Spatiotemporal variability of precipitation isotopes in Colombia inferred with machine learning: Caribbean, Pacific and Andean regions

This Zenodo record contains the supplementary files associated with the manuscript:“Spatiotemporal variability of precipitation isotopes in Colombia inferred with machine learning: Caribbean, Pacific and Andeanregions”. Contents:1) Supplementary_Information.zip - Figure S1: Station-level boxplots used for exploratory QC - Figure S2: Regional distribution of station-wise RMSE - Figure S3: Station-wise RMSE versus elevation - Table S1: Predictor inventory - Table S2: Station-level performance metrics - Table S3: Regional summaries of station-wise RMSE 2) Supplementary_Data_Isoscapes.zip - Monthly isoscapes (GeoTIFF) for δ18O, δ2H and d-excess (dxs), 1971–2024 (0.5°×0.5°) - Supplementary Video S1 (animation) 3) Supplementary_Methods_Scripts.zip - R scripts used to compile predictors, prepare observations, train models, and generate gridded and point-based outputs. - README describing workflow structure and provenance (adapted from the Piso.AI workflow; Nelson et al., 2021). Note: Climate predictors and teleconnection indices used as inputs are publicly available from their official providers; isotope observations were compiled from GNIP (IAEA/WMO) and national/regional sources, preserving station-level provenance. Contact: [e-mail]